Reprogramming of gene expression during compression wood formation in pine

17 Pages·2012·1.75 MB·English

Villalobosetal.BMCPlantBiology2012,12:100
http://www.biomedcentral.com/1471-2229/12/100
RESEARCH ARTICLE Open Access
Reprogramming of gene expression during
compression wood formation in pine: Coordinated
modulation of S-adenosylmethionine, lignin and
lignan related genes
David P Villalobos1,2, Sara M Díaz-Moreno1,3, El-Sayed S Said1, Rafael A Cañas1, Daniel Osuna1,4,
Sonia H E Van Kerckhoven1, Rocío Bautista1, Manuel Gonzalo Claros1, Francisco M Cánovas1 and
Francisco R Cantón1*
Abstract
Background: Transcript profiling of differentiating secondary xylem has allowed us to draw a general picture of the
genes involved in wood formation. However, our knowledge is still limited about the regulatory mechanisms that
coordinate and modulate the different pathways providing substrates during xylogenesis. The development of
compression wood inconifersconstitutes an exceptional model for these studies. Althoughdifferentialexpression
ofa few genes indifferentiating compression wood compared to normal or oppositewood has been reported, the
broad range offeatures that distinguishthis reaction wood suggest thatthe expression of a larger set of genes
would be modified.
Results: Bycombining the construction of different cDNA libraries withmicroarrayanalyses wehave identified a
total of 496genes in maritime pine(Pinuspinaster,Ait.) that change in expression during differentiation of
compression wood (331 up-regulated and 165 down-regulated compared to oppositewood). Samples from
different provenances collectedin different years and geographic locationswere integrated into theanalyses to
mitigate theeffects of multiple sources of variability. This strategy allowed us to define a group of genes that are
consistently associated with compression wood formation.Correlatingwith thedeposition of a thicker secondary
cell wall thatcharacterizescompression wood development, theexpression of a number ofgenes involved in
synthesis of cellulose, hemicellulose, lignin and lignans was up-regulated. Further analysis of a set ofthesegenes
involved in S-adenosylmethionine metabolism, ammonium recycling,and lignin and lignans biosynthesis showed
changes in expression levels in parallel to the levels of lignin accumulationin cells undergoing xylogenesis in vivo
and in vitro.
Conclusions: The comparative transcriptomic analysis reported here have revealed a broad spectrum of
coordinated transcriptional modulation ofgenesinvolved inbiosynthesis of different cell wall polymers associated
withwithin-tree variations inpine wood structure and composition.In particular, wedemonstrate thecoordinated
modulation attranscriptional level ofa gene set involved inS-adenosylmethionine synthesis and ammonium
assimilation with increased demand for coniferyl alcohol for lignin and lignan synthesis, enablinga better
understanding of the metabolic requirements in cells undergoing lignification.
*Correspondence:[email protected]
1DepartamentodeBiologíaMolecularyBioquímica,FacultaddeCiencias,
UniversidaddeMálaga,CampusUniversitariodeTeatinos,29071,Málaga,
Spain
Fulllistofauthorinformationisavailableattheendofthearticle
©2012Villalobosetal.ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttribution
License(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,andreproductioninany
medium,providedtheoriginalworkisproperlycited.
Villalobosetal.BMCPlantBiology2012,12:100 Page2of17
http://www.biomedcentral.com/1471-2229/12/100
Background candidate genes for wood property variations requires
Large amounts of wood can be formed throughout the analysisthatmitigatessourcesofvariationfromgenotypes
lifeofatreethroughacomplexprocessofcelldifferenti- andlocalenvironmentalchanges.
ation called xylogenesis. In this process, cambium- Reaction wood illustrates how the integration of envir-
derived cells undergo cell division followed by thicken- onmental signals into the secondaryxylem developmental
ing ofthe secondarycell wall by modification of the syn- program leads tointra-genotype variation in wood. When
thesis and deposition of cellulose, hemicelluloses, cell thestemofawoodyplantgrowsinanon-verticalorienta-
wall proteins and lignin, and finally programmed cell tionitformsreactionwood,aspecializedsecondaryxylem
death to develop tracheary elements [1]. Genes involved that helps the stem maintain a certain orientation and re-
in these cellular processes are under strict transcrip- orients growth [20]. In gymnosperms, reaction wood is
tional regulationduringdifferent stages ofdifferentiation calledcompressionwoodanditdevelopsontheunderside
[2]. Nevertheless, inputs from external cues are also of branches and inclined stems [21]. Cell walls in com-
integrated in this developmental program to adapt sec- pression wood tracheids are thicker than in normal wood
ondary xylem properties to growth requirements in a andlacktheinnermostS3layer.Italsocontainsmorelig-
continuously changing environment [3]. As a result of nin,lesscelluloseandalteredlevelsof hemicellulosesthan
this interaction, natural variations are found in wood normal wood. Moreover, in compression wood cellulose
properties not only among different species and geno- microfibrils are deposited with increased angle relative to
types,but alsowithin thesame tree[4]. the cell axis [22,23], and the lignin is enriched in p-
Transcriptome analysis of differentiating secondary xylem hydroxyphenylpropane units [24,25]. In contrast, opposite
hasallowedustodrawageneralpictureofthegenes,meta- wood develops on the opposite side of the inclined stem
bolicpathwaysandpotentialregulatorsinvolvedinwoodfor- anditismoresimilartonormalwood[25].Differentialex-
mation[5-11].However,ourknowledgeislimitedastohow pression of a few genes during compression wood differ-
thetranscriptomes,proteomesandmetabolomesinvolvedin entiation compared to normal or opposite wood has been
wooddevelopmentaremodulatedbydevelopmentalanden- reported [16,26-28]. However, the clear anatomical, struc-
vironmental signals to cause within-tree variation in wood turalandcompositionaldifferencesthatcharacterizecom-
properties.Considerableefforthasbeenfocusedonstudying pression wood suggest that the expression of more genes
themainpathwaysthatleadtomonolignolbiosynthesis[12] wouldbemodifiedduringitsdifferentiation[25,29,30].
and carbohydrate partitioning to cellulose [13], as well as We are interested in how biosynthetic pathways that pro-
understanding how changes in gene expression in those videsubstratesforxylogenesisarecoordinatedandregulated
pathways may affect cellular wall characteristics and, conse- according to the different demands during development of
quently,woodquality[14].However,lessattentionhasbeen woodwithdistinctcomposition.Inparticular,comparisonof
paid to the pathways that provide S-adenosylmethionine samplesfromdifferentiatingcompressionwoodandopposite
(SAM)formethylationreactionsduringbiosynthesisofconi- woodfromthesametreeallowsustoanalyzethetranscrip-
feryl and sinapyl alcohols, even though transcripts and pro- tome changes accompanying to different levels of lignin de-
teins for enzymes of these pathways are abundant in the position during xylogenesis. In this work, we present the
developingxylem[15,16].Consumptionofmethylunitsdur- resultsofacomparativetranscriptomicanalysisthatcompre-
ing lignification also implies the existence of an important hensivelyusesarangeofcDNAlibrariesandmicroarrayana-
carbon sink, and modifications in the availability of SAM lyses, combining samples from different Pinus pinaster
mayaffectwoodqualitythroughalterationsinlignincontent provenancescollectedindifferentyearsandgeographicloca-
andcomposition[16-18].Furthermore,ithasbeenproposed tions, to identify changes in the transcriptome associated
thatglycinedecarboxylaseactivityassociatedwithSAMme- withcompressionwooddevelopment.
tabolism could be an important source of released ammo-
nium, which must be efficiently re-assimilated to prevent Results
significant imbalances in the strict nitrogen economy of the Identificationofgenesdifferentiallyexpressedduring
plant[19]. compressionandoppositewoodformationinPinus
Molecular determinants of intra-genotype variation can pinaster
be identified by comparing transcriptomes from cambial Toidentifygenesthataredifferentiallyexpressedbetween
derivativesundergoingdifferentiationinthesametreethat compression(Cx)andopposite(Ox)developingxylemwe
produce wood that differs in composition and structure followed the strategy depicted in Figure 1. Samples from
[4]. However, the plasticity of the molecular machinery different provenances collected in different years at two
involved in wood formation [3] may result in some diffi- latitudes were integrated into the analyses to mitigate the
culties in terms of defining candidate genes when expres- effects of multiple sources of variability. Two microarrays
sion patterns are compared over different annual growth were manufactured with cDNA libraries constructed with
periods. Therefore, the definition of a consistent set of RNA isolated from adult maritime pine trees of Corsican
Villalobosetal.BMCPlantBiology2012,12:100 Page3of17
http://www.biomedcentral.com/1471-2229/12/100
Figure1Schematicoverviewoftheexperimentalprocedureusedinthisworktoidentifygenesdifferentiallyexpressedbetween
compressionandoppositedevelopingxylemfrommaritimepine.Samplesofdifferentiatingsecondaryxylemformingcompression(Cx),
opposite(Ox),early(Ex),late(Lx),juvenile(Jx)andmature(Mx)woodwereusedforcDNAlibrariesconstruction.
and Aquitaine provenances grown in Aquitaine, France. (Figure 2). Consistent with the typical morphology of
The first microarray (Figure 1, cDNA microarray 1) was compression wood tracheids [25], transversal sections of
constructed using a set of 2800 cDNAs from samples of developing tracheids showed a rounded shape and many
Cx, Ox, early (Ex) and late (Lx) developing xylem col- intercellular spaces, whereas Ox developing tracheids had
lected in 1998, 1999 and 2000. The second microarray straight sides and few intercellular spaces (Figure 2a).
(Figure1,cDNAmicroarray2)wasconstructedwith4041 Lignin content was determined in samples from the four
cDNA clones from four subtractive libraries obtained by trees used in microarray analyses and the four trees used
reciprocally subtracting cDNA populations obtained from in qRT-PCR (Figure 2b). Cx samples contained signifi-
Cx and Ox samples that were collected in 1998 and from cantly more lignin than Ox samples. However, variability
juvenile (Jx) and mature (Mx) developing xylem collected in lignin content between both types of developing xylem
in 2001. Both microarrays were hybridized with labeled was evident among trees. In particular, the difference in
aRNAtargetssynthesizedfromCxandOxcollectedfrom lignin contentbetween Cx and Ox samples from treeT4-
four maritime pine trees of the Sierra Bermeja proven- 08waslowercomparedtotheothertreesandnotstatisti-
ance,Spain[31],sampledin2005.Tovalidatetheexpres- callysignificantaccordingtoat-test.
sion patterns of selected genes, northern blot and qRT- Resultsofhybridizationanalysesofbothmicroarraysare
PCRanalyseswerecarriedoutwithsamplesofadulttrees summarizedasvolcanoplotsinFigure3.Aftereliminating
from Sierra Bermeja provenance collected in 2005 and spotsflaggedasbad,notfoundorabsentandthosewitha
2008,respectively. signal intensity that did not surpass twofold their back-
Microscopicexaminationandligninquantificationwere groundsignal,only2598and3148spotswerekeptforfur-
performed for phenotypic validation of the samples ther analysis in microarray 1 and 2 respectively. Relative
Villalobosetal.BMCPlantBiology2012,12:100 Page4of17
http://www.biomedcentral.com/1471-2229/12/100
Figure2Cross-sectionalimagesandlignincontentanalysisofcambialscrapingsamplesusedingeneexpressionanalyses.(a)
Representativemicroscopicimagesofthintransversalcross-sectionsofcambialscrapingsstainedwithphloroglucinol.Theorientationofthe
differentiatingtracheidsinthescrapingrelativetothecambiumanddevelopedxylemisindicatedatthebottomofthepictures.Bluearrows
marktypicalintercellularspacesincompressionwood,andblackarrowsmarkcellswithcharacteristicroundedshapeincompressionwoodor
straightsidesinoppositewood.Scalebaratthebottomleftrepresents0.1mm.(b)Lignincontentofcambialscrapingsfromcompression(black
bars)andopposite(greybars)sidesoftheeightdifferentmaritimepinetreesanalyzed,sampledinMay2005(T1-05,T2-05,T3-05,T4-05)andMay
2008(T1-08,T2-08,T3-08,T4-08)aftertwomonthsofartificialmechanicalbending.Linesonthetopofeverybarshowstandarderrorofthe
meanforfourindependentreplicates.Statisticalsignificanceofdifferencesinlignincontentbetweencambialscrapingsfromcompressionand
oppositesidesfromthesametreewasdeterminedwithapairedStudent´st-test(p<0.05).Alldifferenceswerestatisticallysignificantexceptfor
treeT4-08.
expression values (log Fold Change), p-values and Sequences with an average Phred score >20 per nu-
2
adjusted p-values for all the spots represented in Figure 3 cleotide in a sliding window of 15 nucleotides and a
are available in Additional files 1 and 2. A clear bias is minimal length of 100 nucleotides were kept for further
observed in both cases, with a higher number of spots analysis (average length of 437 nucleotides and a stand-
showing up-regulation in Cx compared to Ox. The ana- ard deviation of ±171). Those with lower quality scores
lysis of microarray 1 shows 229 spots containing probes or shorter than 100 nucleotides were re-sequenced. The
for genes that were significantly up-regulated (p ≤ 0.001) complete set of sequences corresponding to differentially
by at least 1.5-fold during compression wood formation, expressed genes from microarray 1 and 2 were used to
and only 90 spots showing up-regulation during opposite define a set of unigenes, and a total number of 355 uni-
woodformation(Figure3a).SincethesecDNAswerepre- genes up-regulated in Cx and 176 unigenes up-regulated
viously sequenced to construct an EST database of Pinus in Ox were obtained. Moderate values of sequence
pinaster developing xylem the sequences were already redundancy were obtained (31.6% for Cx and 21.2% for
available [32]. The analysis of microarray 2 identified 291 Ox).
spotsshowingup-regulationinCxand132spotsshowing To validate the transcriptomic analysis, a group of
up-regulation in Ox (Figure 3b). Only these 423 cDNA geneswasselectedandtheircDNAswereusedasprobes
clonesweresequencedfromtheprobesetofmicroarray2. for northern blot analysis of RNA extracted from Cx
Villalobosetal.BMCPlantBiology2012,12:100 Page5of17
http://www.biomedcentral.com/1471-2229/12/100
Figure3Volcanoplotsofmicroarrayanalysestoidentifygenes
differentiallyexpressedduringcompressionandopposite
woodformation.Thecommonlogarithmofthep-valuewas
representedasafunctionofthebinarylogarithmofthe
background-correctedandnormalizedopposite:compression
fluorescenceratio(log FoldChange)foreachspot.Verticalbars
2
delimitthespotsshowingup-regulationindevelopingcompression
xylembyatleast1.5-foldcomparedtodevelopingoppositexylem
(Up-regulatedinCx)orspotsshowingup-regulationindeveloping
oppositexylembyatleast1.5-foldcomparedtodeveloping
compressionxylem(Up-regulatedinOx).Thehorizontallinedelimits
thespotsshowingsignificantup-regulationunderthecriteriaofan
adjustedp-value≤0.001.Therefore,theupperleftandrightsectors
delimitedbythehorizontalandverticallinesincludethespots(in
red)containingprobesforgenessignificantlyup-regulatedin
developingcompressionoroppositexylemrespectively.The
numberofspotscorrespondingwithgenessignificantlyup-
regulatedinCxorOxareshowninthetopsideofthesector.(a)
Resultsfromtheanalysisofmicroarray1constructedwithcDNA
clonesfromthecompositelibrary.(b)Resultsfromtheanalysisof
microarray2constructedwithcDNAclonesfromsubtractive
libraries.
using the Pine Gene Index database (Additional file 3).
Sequencesthatmatched withthesame entryinthedata-
base were assumed to represent the same gene. There-
fore, the final numbers of unigenes were reduced to 331
for Cx and 165 for Ox. Most of these genes showed sig-
nificant similarities to sequences in databases (293 in Cx
and 145 in Ox), although some of them were similar to
sequences with unknown function (49 in Cx and 45 in
Ox). The number of unigenes with no significant simi-
larity waslowinbothcases(38inCxand20 inOx).
The genes with assigned function were grouped into
functionalcategoriesusingtheArabidopsisthalianaMun-
ich Information Centre for Protein Sequences (MIPS)
database, and suppression of redundancy in MIPS funcat
assignations by decision according to their most probable
role in xylem development (Additional file 3). In keeping
with the greater number of genes identified as up-
and Ox (Figure 4). The differential expression patterns regulatedinCx,mostofthefunctionalcategoriesincluded
in Cx and Ox observed in microarrays analyses were more genes in this tissue (Figure 5 and Additional file 3
validated for all selected genes. In particular, northern for details). The most represented categories were “C-
blot analysis suggested that our strategy allowed us to compound and carbohydrate metabolism” (30 in Cx/9 in
identify some genes that may be specifically induced at Ox), “cellular transport” (25 in Cx/8 in Ox), “protein syn-
high levels inCx whereas are not expressed or expressed thesis” (18 in Cx/12 in Ox), “amino acid metabolism” (26
at very low levels in Ox (Figure 4, EFE, GST, GLP, XGT, in Cx/2 in Ox), “secondary metabolism” (24 in Cx/2 in
LHT). As a control, the expression pattern was also con- Ox), “cell rescue, defense and virulence” (11 in Cx/15 in
firmed for a gene that is expressed at similar levels in Ox) “protein fate” (14 in Cx/11 in Ox), “biogenesis of cell
both types of differentiating xylem according to micro- wallcomponents”(13inCx/8inOx),“energy”(17inCx/0
arrayanalysis(Figure 4b,panel labeled 13-7XLA6). in Ox), “biogenesis of cytoskeleton” (10 in Cx/2 in Ox),
and“cellularcommunication/signaltransductionmechan-
Trancriptomechangesinfunctionalcategoriesrelatedto ism”(8inCx/3inOx).
cellwallduringcompressionwoodformationinpine The functional categories with larger numbers of up-
The complete set ofunigeneswasfunctionallyannotated regulated genes in Cxcompared to Ox are consistent with
using BlastX analysis [33] against GenBank and BlastN structural and chemical modifications of the cell wall, and
Villalobosetal.BMCPlantBiology2012,12:100 Page6of17
http://www.biomedcentral.com/1471-2229/12/100
particular,thegeneencoding xyloglucangalactosyltransfer-
ase(XGT)showedahighlevelofexpressioninCx,whereas
transcriptswereundetectableinOx(Figure4b,XGT).
Transcripts for enzymes involved inligninbiosynthesiswere
alsoup-regulatedinCx(Additionalfile3,“Up-regulatedinCx”
tab),includingthoseencodingenzymesinthemonolignolbio-
syntheticpathway,suchasphenylalanineammonialyase(PAL:
contig3S,singletonsBX249569,BX253670andFN257113),p-
coumarate 3-hydroxylase (C3H: singleton BX248886), 4-
coumarate:CoA ligase (4CL: singletons BX255483 and
BX253310), cinnamyl alcohol dehydrogenase (CAD: contig 2P,
singletonBX253572),hydroxycinnamoyl-CoA:shikimatehydro-
xycinnamoyl transferase (HCT, singleton FN256636), caffeoyl-
CoA-O-methyltransferase (CCoAOMT, singleton FN256577)
andcaffeateO-methyltransferase(COMT,contig19P).
Several genes up-regulated in Cx and classified in the
functional category “amino acid metabolism” encode
Figure4Validationoftheexpressionprofilesofselectedgenes enzymes involved in the synthesis of precursors for mono-
incompressionandoppositedevelopingxylem.Equalamounts lignol biosynthesis, such as the enzymes of the main trunk
oftotalRNAisolatedfromcompression(Cx)andopposite(Ox) of the shikimate pathway chorismate synthase (contig 5,
differentiatingxylemfromadultmaritimepinetreesweresize
singleton FN256628), 3-dehydroquinate dehydratase/shiki-
fractionedbyelectrophoresisandtransferredtonylonmembranes.
mate5-dehydrogenase(contig22),shikimatekinase(single-
PCR-amplifiedcDNAfrom21genesidentifiedasdifferentially
expressedinanalysesofcDNAmicroarray1(a)orcDNAmicroarray ton BX251461) and 5-enolpyruvylshikimate 3-phosphate
2(b)werelabeledwith32Pandusedasprobes.Thepanel13-7XL synthase(singletonBX252482).Inparticular,theexpression
A6showstheresultofhybridizationwithcDNAfromagene pattern was confirmed by northern blot for the genes en-
classifiedasnon-differentiallyexpressedbythemicroarrayanalyses.
coding chorismate synthase, and 3-dehydroquinate dehy-
Panelsbeloweverynorthernblotshowmethylenebluestaining,as
dratase/shikimate 5-dehydrogenase (Figure 4a, CS and
aloadingcontrol.Functionalannotationandaccessionnumberof
thecorrespondingESTs:CS(chorismatesynthase,BX250851),DHQ/ DHQ/SDH,respectively).
SDH(3-dehydroquinatedehydratase/shikimate5-dehydrogenase,
BX251808),mSHMT(mitocondrialserinehydroxymethyltransferase, ExpressionofgenesencodingenzymesrelatedtoSAM
BX249820),EFE(ethylene-formingenzyme,BX251740),LAC(laccase,
biosynthesisincellsundergoingxylogenesis
BX252150),GST(glutathioneS-transferase,BX250376),MS
Genes related to S-adenosylmethionine (SAM) metabol-
(methioninesynthase,BX255477),SAMS(S-adenosylmethionine
synthetase,BX248790),GS1b(glutaminesynthetase1b,BX253698), ism were also widely represented among those up-
GLP(germin-likeprotein,FN256510),XGT(xyloglucan regulated in Cx and classified in the functional category
galactosyltransferase,FN256497),LHT(lysine/histidinetransporter, “aminoacidmetabolism” (Additional file3,“Up-regulated
FN256535),PLR(pinoresinol-lariciresinolreductase,FN256693),GDCH in Cx” tab). SAM is a universal methyl donor for many
(glycinedecarboxylasecomplexH-protein,FN256878),COMT
different cellular metabolic reactions, including those cat-
(caffeateO-methyltransferase,FN256925),CCoAOMT(caffeoyl-CoA-O-
methyltransferase,FN256577),cSHMT(cytosolicserine alyzed by CCoAOMTand COMT. Both enzymes are es-
hydroxymethyltransferase,FN256831),SAHH(S-adenosyl sential for the biosynthesis of coniferyl and sinapyl
homocysteinehydrolase,FN256791),unknown(unknown-function alcohols, precursors of lignin, lignans and other phenyl-
gene,FN256975),XET(xyloglucanendotransglucosylase/hydrolase,
propanoids [34]. Therefore, the higher lignin content of
FN256785),13-7XLA6(cDNAofagenewiththesamelevelof
compression wood may require a larger supply of SAM
expressioninCxandOxaccordingtomicroarrayanalyses).
duringdevelopmentcomparedtooppositewood.
To unambiguously validate the differential expression
asshowninAdditionalfile3(“Up-regulatedinCx”tab)in- patterns observed in the microarray analyses for genes
clude genes encoding cellulose synthase subunits (contigs related to SAM metabolism, relative transcript levels were
19 and 7S, singletons BX249614 and FN256963), sucrose analyzed by qRT-PCR in Cx and Ox of four independent
synthase (contigs 12 and 35S), cellulose synthase-like A trees from Sierra Bermeja that were sampled in 2008 (Fig-
(contig 19S), glycosyltransferases (contig 4P, singleton ure 6 shows fold change values of relative transcript abun-
BX250177), xyloglucan galactosyltransferase (singleton dance and values for normalized relative abundance are
FN256497),α-andβ-tubulins(contigs13,7P,31Sand51S, provided in Additional file 4). These genes encode the
singletons BX255503, FN256632, FN256674, FN256784, enzymesoftheactivatedmethylcycleandincludedmethio-
BX255285, BX249611 and FN256695) and a putative nine synthase (MS: contigs 27 and 21S, singletons
microtubule-associated protein (singleton BX253390). In BX255406, BX255477, BX251773, FN257093 and
Villalobosetal.BMCPlantBiology2012,12:100 Page7of17
http://www.biomedcentral.com/1471-2229/12/100
Figure5ThemostrepresentedfunctionalcategoriesamongthegenesdifferentiallyexpressedinCxandOx.Thecompletesetof
unigeneswasgroupedintofunctionalcategoriesusingtheMIPSdatabase.Thosecategorieswithatotalnumberofunigenesover10were
consideredasthemostrepresented.Thetotalnumberofunigenesisshownbetweenbracketsfollowingthecategoryname.Thenumberof
unigenesup-regulatedinCxisrepresentedasblackbars,andthenumberofunigenesup-regulatedinOxisrepresentedasgreybars.
FN256480), S-adenosylmethionine synthase (SAMS: single- those genes that were not available in databases (see Add-
tons BX248790, FN257080, FN256463, FN256552, itionalfile5foraccessionnumbers).Allthesegeneswerein-
FN564393) and S-adenosylhomocysteine hydrolase (SAHH: deed up-regulated in Cx compared to Ox in the four trees
contig 22S). The expression of genes encoding enzymes of according to fold change values of relative transcript abun-
the monolignol pathway branch that utilizes methyl groups dance(Figure6andAdditionalfile4).However,itisinterest-
provided by SAM to produce methoxylated monolignols ing to note that smaller differences in transcript levels
were also analyzed, including hydroxycinnamoyl-CoA:shiki- betweenthetwotypesofxylemwereobservedintreeT4-08.
matehydroxycinnamoyltransferase(HCT),Caffeoyl-CoA-O As a control, the relative transcript levels for two reference
-methyltransferase (CCoAOMT) and Caffeate O- genesusedinthenormalizationofqRT-PCRdataareshown
methyltransferase(COMT).Othergenesincludedintheana- inAdditionalfile6.
lysis encode enzymesinvolvedin thesupplyofmethylunits To further demonstrate the relationship between lig-
into the activated methyl cycle from the amino acid serine, nin accumulation and enhanced expression of genes
such as cytosolic methylenetetrahydrofolate reductase related to SAM metabolism, we quantified by qRT-PCR
(MTHFR: singletons BX254818 and FN564392) and serine the relative levels of transcripts in cells from Pinus pin-
hydroxymethyltransferase (cSHMT: contig 11 and singleton aster calli that were stimulated to differentiate tracheids
FN256755), mitochondrial glycine decarboxylase complex in vitro. We tested the occurrence of tracheids after 0, 7
H-protein(GDCH:singletonFN256878)andserinehydroxy- and 15 days of transfer to the induction medium (EDM
methyltransferase (mSHMT: singleton BX249820), and the without hormones) by phloroglucinol staining and mi-
plastid enzymes 3-phosphoglycerate dehydrogenase croscopy inspection (Figure 7a). Differentiated tracheids
(PHGDH: singletons BX251805, FN256792) and phospho- were present even in the control medium (EDM with
serineaminotransferase(PSAT:singletonFN257119).Finally, hormones), but the number of differentiated tracheids
we also considered a gene encoding glutamine synthetase was higher in induction medium (data not shown). Con-
(GS1b:singletonBX253698)up-regulatedinCxaccordingto sequently lignin levels were higher in samples from in-
microarray analysis (Additional file 3), which has been pro- duction medium (Figure 7b). Transcript levels of genes
posed to play a key role in re-assimilation of ammonium encoding enzymes involved in SAM metabolism were
released from different process in developing xylem [19,35]. higher in cells transferred to induction medium com-
To unambiguously verify the identity of the encoded pro- pared to those transferred to control medium, in parallel
teins, we cloned by RLM-RACE the full-length cDNAs for with the observed increase in lignin (Figure 7c). In most
Villalobosetal.BMCPlantBiology2012,12:100 Page8of17
http://www.biomedcentral.com/1471-2229/12/100
Figure6EnzymesinvolvedinS-adenosylmethioninemetabolismanddifferentialexpressionofthecorrespondinggenesinCxandOx.
Transcriptlevelswereanalyzedinsamplesfromfouradulttrees(T1-08,T2-08,T3-08,T4-08inFig.2)aftertwomonthsofartificialbendingof
stems,andtheCx/OxfoldchangeineachtreeasdeterminedbyqRT-PCRisrepresentedinfoursquaresforeverygenebyaredcolorscale.
Numericalvaluesoffoldchangerepresentedbythecolorscaleareshownatthebottomright.GDC:glycinedecarboxylasecomplex;mSHMT:
mitochondrialserinehydroxymethyltransferase;cSHMT:cytosolicserinehydroxymethyltransferase;MS:methioninesynthase;MTHFR:
methylenetetrahydrofolatereductase;SAMS:S-adenosylmethioninesynthase;COMT:caffeateO-methyltransferase;CCoAOMT:caffeoyl-CoA-O
-methyltransferase;SAHH:S-adenosylhomocysteinehydrolase;HCT:hydroxycinnamoyl-CoA:shikimatehydroxycinnamoyltransferase;PHGDH:3-
phosphoglyceratedehydrogenase;PSAT:3-phosphoserineaminotransferase;PSP:3-phosphoserinephosphatase;NADHGOGAT:NADH-dependent
glutamatesynthase;GS1b:glutaminesynthetase1b;THF:tetrahydrofolate;HCy:homocysteine;SAHcy:S-adenosylhomocysteine;3PGA:3-
phosphoglycerate;3PHP:3-phosphohydroxypyruvate;3PSer:3-phosphoserine.
of the genes the level of transcripts in cells after 15 days singleton BX255256) are involved in the 8-5’ linked lig-
in induction medium was at least twice the level in cells nans pathway [37]. The differential expression of genes
thatwere cultured incontrol medium. encoding these enzymes was also validated in four trees
by qRT-PCR, with higher transcript levels in Cx com-
pared toOx(Figure 8a).Notably,thisdifferencewasalso
Expressionofgenesencodingenzymesrelatedtolignan- lower in tree T4-08. Transcript levels were also
biosynthesisincellsundergoingxylogenesis increased in cells of Pinus pinaster calli that were
In addition to being a lignin precursor, coniferyl alcohol induced to differentiate into tracheids compared to non-
is also a precursor of different lignans [36]. Three genes inducedcells(Figure8b).
encoding enzymes involved in biosynthesis of coniferyl
alcohol-derived lignans were also found among the up-
regulated genes in Cx, according to microarray analyses Discussion
(Additional file 3). Pinoresinol-lariciresinol reductase Comparativetranscriptomeanalysisshows
(PLR: contig 44S, singletons BX251658 and BX252917) reprogrammingofcellwallrelatedgenesduring
isanenzymeinvolvedinthe8-8’linked lignanspathway, compressionwoodformationinpine
and phenylcoumaran benzylic ether reductase (PCBER: In this work, we carried out a comparative transcriptome
contigs 17and17S,singleton FN256729)andphenylpro- analysis between developing compression and opposite
penal double-bond reductase (PPDBR: contig 36S, woodusingacombinationofdifferentlyconstructedcDNA
Villalobosetal.BMCPlantBiology2012,12:100 Page9of17
http://www.biomedcentral.com/1471-2229/12/100
Figure7LigninaccumulationandexpressionofgenesinvolvedinS-adenosylmethioninemetabolismduringinvitrotracheid
differentiationfromPinuspinastercalluscells.(a)MicroscopicimageofPinuspinastertracheidsdevelopedinvitro.Lignindepositedina
typicalbandpatternwasstainedwithphloroglucinol.Scalebaronthebottomleftrepresents10μm.(b)LignincontentsincultureofPinus
pinastercellsafter0,7and15daysoftransfertoinductionmedium(withouthormones,opencircles)orcontrolmedium(withhormones,closed
circles).Valuesarethemeansofthreeindependentquantifications±SE.(c)RelativetranscriptabundanceofgenesinvolvedinS-
adenosylmethioninemetabolismafter0,7and15daysoftransferringPinuspinastercalluscellstoinductionmedium(withouthormones,open
circles)orcontrolmedium(withhormones,closedcircles).Valuesarethemeansofthreeindependentreplicates±SE.
Villalobosetal.BMCPlantBiology2012,12:100 Page10of17
http://www.biomedcentral.com/1471-2229/12/100
Figure8Relativetranscriptabundanceofthreegenesinvolvedinlignanbiosynthesisincellsundergoingxylogenesis.Relative
transcriptabundanceforeachgenewasdeterminedbyqRT-PCR.(a)Relativetranscriptlevelsincompression(blackbars)andopposite(greybars)
developingxylemfromfouradulttrees(T1-08,T2-08,T3-08,T4-08)aftertwomonthsofartificialbendingofthestems.(b)Relativetranscript
levelsafter0,7and15daysoftransferringPinuspinastercalluscellstoinductionmedium(withouthormones,opencircles)orcontrolmedium
(withhormones,closedcircles).Valuesaremeans±SEofthreeindependentreplicates.PLR:Pinoresinol-lariciresinolreductase;PCBER:
phenylcoumaranbenzylicetherreductase;PPDBR:phenylpropenaldouble-bondreductase.
libraries and microarray analyses (Figure 1). Samples from The external stimulus that leads to the formation of
three unrelated provenances, different geographic location compression wood caused significant changes of gene ex-
or collected in different years (Figure 2) were integrated in pression, as revealed by the large number of genes that
the analysis to mitigate the potential effects of variability were up-regulated or down-regulated in Cx compared to
triggered by genotypic or environmental differences in Ox(Figure3)andthespecificexpressionofsomegenesin
order to identify a consistent set of candidate genes for Cx (Figure 4). An important component of this gene re-
woodproperties. programming is clearly related to the chemical and